Tetrahedral and square planar Ni[(SPR2)2N] 2 complexes, R = Ph &iPr revisited: Experimental and theoretical analysis of interconversion pathways, structural preferences, and spin delocalization

Article abstract of DOI:10.1021/ic100163g

Sulfur-containing mono-or bidentate types of ligands, usually form square planar Ni^(II)S₄ complexes. However, it has already been established that the bidentate L^- dithioimidodiphosphinato ligands, [R₂P(S)NP(S)R′₂]^-, R, andR′ = aryl or alkyl, can afford both tetrahedral and square planar, NiS₄- containing, homoleptic Ni^R,R′L₂ complexes, owing to an apparent structural flexibility, which has not, so far, been probed. In this work, the literature tetrahedral Ni[R₂P(S)NP(S)R₂] ₂ complexes, R = Ph (Ni^Ph,PhL₂, 1_Td) and R = ⁱPr (Ni^iPr,iPrL₂, 2) as well as the newly synthesized Ni[ⁱPr₂P(S)NP(S)Ph₂] ₂ complex (Ni^iPr,PhL₂, 3), have been studied by UV-vis, IR, and ³¹P NMR spectroscopy. Complex 3 was shown by X-ray crystallography to be square planar, and magnetic studies confirmed that it is diamagnetic in the solid state. However, it becomes paramagnetic in solution, as it shows a similar UV-vis spectrum to one of the tetrahedral 1_Td and 2 complexes. The crystal structure of the potassium salt of the asymmetric ligand, [ⁱPr₂P(S)NP(S)Ph₂]K, has also been determined and compared to those of the protonated ⁱPr ₂P(S)NHP(S)Ph₂ ligand and complex 3. All three, 1 _Td, 2, and3, Ni^R,R′L₂ complexes show strong paramagnetic effects in their solution ³¹P NMRspectra. Themagnetic properties of paramagnetic complexes 1 and 2 in the solid state were investigated on oriented crystals, and their analysis afforded remarkably small values of the spin-orbit coupling constant (λ) and orbital reduction factor (k) parameters, implying significant delocalization of unpaired electronic density toward the ligands. The above experimental findings are combined with data from standard density functional theory and correlated multiconfiguration ab initio theoretical methods, in an effort to investigate the interplay between the square planar and tetrahedral geometries of the NiS₄ core, the mechanistic pathway for the spin-state interconversion, the degree of covalency of the Ni-S bonds, and the distribution of the spin density in this type of system. The analysis provides justification for the structural flexibility of such ligands, affording Ni ^R,R′L₂ complexes with variable metallacycle conformation and NiS₄ core geometries. Of particular importance are the large zero-field splitting values estimated by both experimental and theoretical means, which have not, as yet, been verified by direct methods, such as electron paramagnetic resonance spectroscopy. The findings of our work confirm earlier observations on the feasibility of synthesizing either tetrahedral or square planar NiS4 complexes containing the same type of ligands. They can also form the basis of investigating structure-properties relationships in other NiS₄-containing systems.

Full text of DOI:10.1021/ic100163g

Inorg. Chem. 2010, 49, 5079–5093 5079
DOI: 10.1021/ic100163g
Tetrahedral and Square Planar Ni[(SPR2)2N]2 complexes, R = Ph & iPr Revisited:
Experimental and Theoretical Analysis of Interconversion Pathways, Structural
Preferences, and Spin Delocalization
Dimitrios Maganas,†,|| Alexios Grigoropoulos,† Sarah S. Staniland,‡,# Spyros D. Chatziefthimiou,§
Andrew Harrison,‡ Neil Robertson,*,‡ Panayotis Kyritsis,*,† and Frank Neese*,||
†Inorganic Chemistry Laboratory, Department of Chemistry, National and Kapodistrian University of Athens, GR-157
71 Athens, Greece, ‡School of Chemistry and EaStChem, University of Edinburgh, West Mains Road, Edinburgh EH9
3JJ, United Kingdom, §EMBL-Hamburg c/o Deutsches Elektronen-Synchrotron, Hamburg, Germany, and
||Institute of Theoretical and Physical Chemistry, Wegelerstrasse 12, D-53115 Bonn, Germany. # Present address:
School of Physics and Astronomy, University of Leeds, Woodhouse Lane, Leeds LS2 9JT, United Kingdom.
Received January 27, 2010
Sulfur-containing mono- or bidentate types of ligands, usually form square planar Ni(II)S4 complexes. However, it has
already been established that the bidentate L- dithioimidodiphosphinato ligands, [R2P(S)NP(S)R02]-, R, and R0 = aryl or
0
alkyl, can afford both tetrahedral and square planar, NiS4-containing, homoleptic NiR,R L2 complexes, owing to an apparent
structural flexibility, which has not, so far, been probed. In this work, the literature tetrahedral Ni[R2P(S)NP(S)R2]2
complexes, R = Ph (NiPh,PhL2, 1Td) and R = iPr (NiiPr,iPrL2, 2) as well as the newly synthesized Ni[iPr2P(S)NP(S)Ph2]2
complex (NiiPr,PhL2, 3), have been studied by UV-vis, IR, and 31P NMR spectroscopy. Complex 3 was shown by X-ray
crystallography to be square planar, and magnetic studies confirmed that it is diamagnetic in the solid state. However, it
becomes paramagnetic in solution, as it shows a similar UV-vis spectrum to one of the tetrahedral 1Td and 2 complexes.
The crystal structure of the potassium salt of the asymmetric ligand, [iPr2P(S)NP(S)Ph2]K, has also been determined and
0
compared to those of the protonated iPr2P(S)NHP(S)Ph2 ligand and complex 3. All three, 1Td, 2, and3, NiR,R L2 complexes
show strong paramagnetic effects in their solution 31P NMR spectra. The magnetic properties of paramagnetic complexes 1
and 2 in the solid state were investigated on oriented crystals, and their analysis afforded remarkably small values of the
spin-orbit coupling constant (λ) and orbital reduction factor (k) parameters, implying significant delocalization of unpaired
electronic density toward the ligands. The above experimental findings are combined with data from standard density
functional theory and correlated multiconfiguration ab initio theoretical methods, in an effort to investigate the interplay
between the square planar and tetrahedral geometries of the NiS4 core, the mechanistic pathway for the spin-state
interconversion, the degree of covalency of the Ni-S bonds, and the distribution of the spin density in this type of system.
0
The analysis provides justification for the structural flexibility of such ligands, affording NiR,R L2 complexes with variable
metallacycle conformation and NiS4 core geometries. Of particular importance are the large zero-field splitting values
estimated by both experimental and theoretical means, which have not, as yet, been verified by direct methods, such as
electron paramagnetic resonance spectroscopy. The findings of our work confirm earlier observations on the feasibility of
synthesizing either tetrahedral or square planar NiS4 complexes containing the same type of ligands. They can also form
the basis of investigating structure-properties relationships in other NiS4-containing systems.
Introduction
experimental and theoretical investigations have established
that the electronic properties of these metalloproteins are
highly dependent on the contribution of the sulfur valence
orbitals to the “redox active” molecular orbitals of the active
site, whichis a measureofthe covalencyofthe M-S bonds.3,4
Crystallographic studies in the past few years have estab-
lished that the active site of a range of nickel enzymes show
A common feature in structural bioinorganic chemistry is
the presence of sulfur-containing ligands in the active site of
metalloproteins and metalloenzymes, most usually in the
form of cysteine thiolate, S(Cys), or sulfide, S2- 1,2
Recent
.
*To whom correspondence should be addressed. E-mail: Kyritsis@
chem.uoa.gr (P.K.), neil.robertson@ed.ac.uk (N.R.), neese@thch.uni-bonn.
de (F.N.).
(1) Holm, R. H.; Kennepohl, P.; Solomon, E. I. Chem. Rev. 1996, 96,
2239–2314.
(2) Rees, D. C.; Howard, J. B. Science 2003, 300, 929–931.
(3) Solomon, E. I.; Gorelsky, S. I.; Dey, A. J. Comput. Chem. 2006, 27,
1415–1428.
(4) Solomon, E. I. Inorg. Chem. 2006, 45, 8012–8025.
r
2010 American Chemical Society
Published on Web 05/12/2010
pubs.acs.org/IC

5080 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
remarkable variation in nickel coordination, and although
the presence of Ni-S bonds in their active sites is rather
prominent, they do not contain intact NiS4 cores.5 However,
an artificial intact [Ni(II)S(Cys)4] center has beengeneratedby
replacing Fe(II) by Ni(II) in the active site of rubredoxin. The
Ni-substituted rubredoxin has been studied by X-ray
crystallography6 and spectroscopic methods,7,8 with an aim
of elucidating the structural and electronic properties of the
metal site. In this case, the protein scaffold provides a rigid
structural framework, imposing a tetrahedral (Td) Ni(II)S4
metal core, rather similar to the structure of the Fe(II)S4 site in
native rubredoxins.6,8 On the other hand, extensive inorganic
synthetic work has shown that Ni(II)S4-containing complexes
usually exhibit square planar (SP) geometry, whereas the Td
geometry is rather rare.9 It should be stressed that the
interchange between SP and Td geometries in NiS4-contain-
ing complexes is not yet fully elucidated, despite promising
theoretical efforts concerning tetra-coordinate transition-
metal complexes.10
effects that favor the SP or Td geometry in NiS4 systems,
given that, depending on the nature of the peripheral R
groups, both geometries have been observed in such com-
plexes under specific experimental conditions. More specifi-
cally, the Td geometry is preferred when the ligand is
symmetric, i.e., both P atoms bear the same peripheral groups
(R = R0), as it can be deduced from the crystal structures of
NiMe,MeL2,23 NiPh,PhL2 (1Td),9,24 and NiiPr,iPrL2 (2).25 The
structure of NinBu,tBuL2, in which each P atom is bonded to
n
t
one Bu and one Bu group, has also revealed a Td NiS4
core.26 It should be noted that, based on spectroscopic data,
NiPh,PhL2 was the first Ni(II)S4 complex reported to adopt a
Td geometry.27 On the other hand, the complexes with
i
asymmetric ligands (R = Ph and R0 = Me or Pr) contain
a SP NiS4 core: the structure of NiMe,PhL2 has already been
reported,9 whereas the synthesis and structural characteriza-
tion of NiiPr,PhL2 (3) is described herein.
The only exemption to the above general trend is
NiPh,PhL2, which has been also crystallized in a form contain-
ing a SP NiS4 core (1SP) and THF molecules.28 To the best of
our knowledge, NiPh,PhL2 is the only NiS4 complex that has
been reported, up to date, to adopt, under different conditions,
both a Td and a SP geometry in the solid state. Such Ni(II)
systems are expected to exhibit not only stereoisomerism but
also spin isomerism, as the Td complexes are typically para-
magnetic (S=1), whereas the SP ones are usually diamagnetic
(S=0),29 although some exceptions have been documented
by the identification of rare paramagnetic SP Ni(II) com-
plexes.30,31 In that respect, it is of interest to note that the
ability of the dichalcogenated imidodiphosphinato ligands to
form stereoisomers upon coordination to Ni(II) might be a
general trend, since the corresponding Ni[iPr2P(Se)NP-
(Se)iPr2]2 complex also affords both Td and SP polymorphs.32
With the aim of advancing our previous investigations on
A search in the Cambridge Structural Database11 reveals
that only a handful out of over 800 complexes bearing the
Ni(II)S4 core are not SP but instead display Td geometry. The
mononuclear Td Ni(II)S4 complexes, that have been structu-
rally characterizedup to date, can be dividedinto two groups:
(i) Ni(II) complexes bearing monodentate arylthiolato ligands
of the [Ni(SAr)4]2- general type,12-18 and (ii) Ni(II) com-
plexes bearing bidentate dithioimidodiphosphinato ligands,
Ni[R2P(S)NP(S)R02]2, R, R0 = aryl or alkyl (denoted as
0
NiR,R L2). The latter type of ligands are considered to be
inorganic analogues ofβ-diketonates, and they readily form a
wealth of M(II)L2 coordination compounds, containing six-
membered chelating rings, that exhibit extensive delocaliza-
tion of π-electronic density among the S-P-N-P-S
fragment.19-22 Furthermore, their wide S S “bite” range
3 3 3
(ca. 4 A) allows them to satisfy the stereochemical require-
analogous complexes of Mn(II) and Co(II) 33,34
, we sought to
˚
ments of different geometries.
explore the structural preferences and the electronic proper-
0
0
The family of the NiR,R L2 complexes offers significant
advantages in elucidating the electronic and stereochemical
ties of NiR,R L2 complexes, both in the solid state and in
solution, by employing magnetization, spectroscopic, and
standard density functional theory (DFT) and correlated
multiconfiguration ab initio theoretical methods. The ques-
tions addressed in this work are: (i) the interplay between SP
and Td geometries of the NiS4 core, depending on the nature
of the peripheral R groups as well as the mechanistic pathway
(5) Ragsdale, S. W. J. Biol. Chem. 2009, 284, 18571–18575.
(6) Maher, M.; Cross, M.; Wilce, M. C. J.; Guss, J. M.; Wedd, A. G. Acta
Crystallogr. 2004, D60, 298–303.
(7) Moura, I.; Teixeira, M.; Legall, J.; Moura, J. J. G. J. Inorg. Biochem.
1991, 44, 127–139.
(8) Goodfellow, B. J.; Duarte, I. C. N.; A. L. Macedo, A. L.; Volkman,
B. F.; Nunes, S. G.; Moura, I.; Markley, J. L; Moura, J. J. G. J. Biol. Inorg.
Chem. 2010, 15, 409–420.
(9) Rosler, R.; Silvestru, C.; EspinosaPerez, G.; Haiduc, I.; CeaOlivares,
R. Inorg. Chim. Acta 1996, 241, 47–54.
(10) Cirera, J.; Alemany, P.; Alvarez, S. Chem.—Eur. J. 2004, 10, 190–
207.
(11) Allen, F. H. Acta Crystallogr. 2002, B58, 380–388.
(12) Yamamura, T.; Miyamae, H.; Katayama, Y.; Sasaki, Y. Chem. Lett.
1985, 269–272.
(13) Rheingold, A. L.; Beall, K. S.; Riggs, P. J.; Groh, S. E. Acta
Crystallogr. 1993, C49, 542–543.
(14) Rosenfield, S. G.; Armstrong, W. H.; Mascharak, P. K. Inorg. Chem.
1986, 25, 3014–3018.
(15) Silver, A.; Koch, S. A.; Millar, M. Inorg. Chim. Acta 1993, 205, 9–14.
(16) Swenson, D.; Baenziger, N. C.; Coucouvanis, D. J. Am. Chem. Soc.
1978, 100, 1932–1934.
(23) Churchil, M. R.; Cooke, J.; Fennessey, J. P.; Wormald, J. Inorg.
Chem. 1971, 10, 1031–1035.
(24) Bhattacharyya, P.; Novosad, J.; Phillips, J.; Slawin, A. M. Z.;
Williams, D. J.; Woollins, J. D. J. Chem. Soc., Dalton Trans. 1995, 1607–
1613.
(25) Cupertino, D.; Keyte, R.; Slawin, A. M. Z.; Williams, D. J.;
Woollins, J. D. Inorg. Chem. 1996, 35, 2695–2697.
(26) Moore, P. Structure deposited in the Cambridge Structural Database
as Personal Communication.
(27) Davison, A.; Switkes, E. S. Inorg. Chem. 1971, 10, 837–842.
(28) Simon-Manso, E.; Valderrama, M.; Boys, D. Inorg. Chem. 2001, 40,
3647–3649.
(29) Cirera, J.; Ruiz, E.; Alvarez, S. Inorg. Chem. 2008, 47, 2871–2889.
(30) Frommel, T.; Peters, W.; Wunderlich, H.; Kuchen, W. Angew.
Chem., Int. Ed. 1993, 32, 907–909.
(31) Bridgeman, A. J. Dalton Trans. 2008, 1989–1992.
(32) Levesanos, N.; Robertson, S. D.; Maganas, D.; Raptopoulou, C. P.;
Terzis, A.; Kyritsis, P.; Chivers, T. Inorg. Chem. 2008, 47, 2949–2951.
(33) Maganas, D.; Staniland, S. S.; Grigoropoulos, A.; F. White, S. P.;
Robertson, N.; Kyritsis, P.; Pneumatikakis, G. Dalton Trans. 2006, 2301–2315.
(34) (a) Maganas, D.; Milikisyants, S.; Rijnbeek, J. M. A.; Sottini, S.;
Levesanos, N.; Kyritsis, P.; Groenen, E. J. J. Inorg. Chem. 2010, 49, 595–605.
(b) Sottini, S.; Mathies, G.; Gast, P.; Maganas, D.; Kyritsis, P.; Groenen, E. J. J.
Phys. Chem. Chem. Phys. 2009, 11, 6727–6732.
(17) Kockerling, M.; Henkel, G. Z. Kristallogr. 1994, 209, 186–187.
(18) Muller, A.; Henkel, G. Z. Naturforsch. 1995, 50b, 1464–1468.
(19) Woollins, J. D. J. Chem. Soc., Dalton Trans. 1996, 2893–2901.
(20) Ly, T. Q.; Woollins, J. D. Coord. Chem. Rev. 1998, 176, 451–481.
(21) Silvestru, C.; Drake, J. E. Coord. Chem. Rev. 2001, 223, 117–216.
(22) Haiduc, I. Dichalcogenoimidodiphosphinato Ligands. In Compre-
hensive Coordination Chemistry II: From Biology to Nanotechnology;
McCleverty, J. A., Meyer, T. J., Eds.; Elsevier: Amsterdam, 2003, Vol. 1,
pp 323-347.

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5081
for the spin-state interconversion, (ii) the degree of covalency
of the Ni-S bonds, depending on the NiS4 geometry, and
(iii) the distribution of the spin density in the paramagnetic
Td complexes.
Computational Details. All electronic structure calculations
reported in this paper were performed with the ORCA
program.39 For the DFT and the ab initio multiconfigurational
self-consistent field theory (MCSCF) calculations, truncated
models denoted as NiH,HL2 in different spin states and con-
formations, as defined in the respective sections, were chosen,
such as to keep the computational effort acceptable. For the
DFT calculations, the BP8640,41 and B3LYP40,42,43 functio-
nals were used for the calculation of geometries/frequencies
and spectroscopic properties, respectively, together with a basis
set of polarized triple-ζ (TZVP) quality. In order to compute the
d-d transitions of Ni(II) complexes, the minimal active space
should include the five 3d-based molecular orbitals involved in
nonbonding or σ M-S antibonding interactions as well as the
corresponding σ M-S bonding molecular orbitals. This leads to
an active space with 12 electrons in 7 orbitals (CAS (12,7)). We
have included all 10 roots for the triplet states (arising from the
3F and 3P terms of Ni2þ) and 5 singlet roots (arising from the 1D
term of Ni2þ). MR-difference dedicated CI calculations, with
two degrees of freedom (DDCI2)44 were performed on top of the
state-averaged-complete active space self-consistent field (SA-
CASSCF) reference wave function, in order to recover the major
part of the differential dynamic correlation between the ground
and excited states. As it was commented upon previously,45 we
have used individual selection in order to decrease the computa-
tion time. The size of the first-order interacting space was
reduced with a threshold Tsel = 10-6 Eh. A further approxima-
tion involves reduction of the reference space through rejec-
tion of all initial references that contribute less than a second
threshold (Tpre = 10-5) to the zeroth-order states. The minimum-
energy crossing point (MECP) of the DFT S = 0 and 1 surfaces
was identified via MECP optimizations, as implemented in the
ORCA program.
Experimental Section
Materials and Methods. All experiments were carried out
under an atmosphere of argon using Schlenk techniques. The
glassware was dried in the oven at approximately 110 ꢀC and
baked out in vacuo prior to use. The solvents were dried by
standard methods (methanol and dichloromethane over CaH2,
n-hexane over sodium wire) and distilled under argon. The
solvents were deoxygenated by at least three pump and purge
cycles immediately prior to use. All chemical reagents were
purchased from Aldrich. The LH ligands, R2P(S)NHP(S)R02
R,R0
(denoted as
LH), RdR0dPh,35 RdR0diPr,25 RdPh,0 and
R0diPr36 and their corresponding LK salts (denoted as R,R LK)
as well as the complexes 1Td
according to published procedures.
9,24
,
28
,
1SP
and 225 were prepared
UV-vis, IR, and NMR Spectroscopy. Electronic spectra were
recorded in a Carry 300 Varian spectrophotometer. IR spectra
were run in the range of 4000-200 cm-1 on a Perkin-Elmer
1
883 IR spectrophotometer, as KBR pellets. H and 31P-{1H}
spectra were recorded in a Varian Unity Plus 300 MHz instru-
ment. The 1H and 31P-{1H} chemical shifts are relative to SiMe4
and to 85% H3PO4, respectively.
Magnetic Susceptibility Studies. Magnetic susceptibility mea-
surements were carried out between 2 and 300 K, using a
Quantum design MPMS2 SQUID magnetometer with MPMS
MultiVu Application software to process the data. The mag-
netic field employed was 0.1 T. The program used for the
simulation of magnetic data is presented in the Supporting
Information.
Synthesis of K[iPr2P(S)NP(S)Ph2] (iPr,PhLK). In a Schlenk-
type flask charged with 5 g (12.1 mmol) of iPr,PhLH, 20 mL of
MeOH are added under stirring. To the resulting suspension, 1.8 g
X-ray Crystallography. Ligand iPr,PhLK. : Low-temperature
X-ray data were collected at the synchrotron radiation light
source at EMBL-Hamburg, Deutsches Elektronen-Synchrotron
(DESY), beamline X13, by the oscillation method using a CCD
detector of 165 mm radius. A single crystal, covered with a drop of
oil, was mounted on a small loop of hair fiber and was instantly
frozen at 100 K. The programs DENZO and SCALEPACK37
were used for data processing and scaling of data, respectively.
The structure was solved by direct methods and refined by
full-matrix least-squares based on F2, using the program
SHELXL97.38 Hydrogen (H) atoms have been placed in idealized
positions and refined by the riding model (UH = 1.30 UC).
Complex 3. : X-ray diffraction intensities were collected with
Mo KR radiation on a Bruker Smart APEX CCD diffractometer
equipped with an Oxford Cryosystems low-temperature device
operating at 150 K. Absorption corrections were carried out with
the multiscan procedure SADABS. The structure was solved by
direct methods (SHELXS97) and refined by full-matrix least-
squares against |F|2 using all data (SHELXL97). H atoms were
placed in calculated positions and allowed to ride on their parent
atoms. All non-H atoms were modeled with anisotropic displa-
cement parameters.
t
of BuOK are added at once. The white mixture is left stirring
for 2 h. At this point, the white solid is filtered and washed
with 3 ꢀ 10 mL of cold MeOH to yield 5.1 g (11.2 mmol) of
iPr,PhLK (93%). White needle-shaped crystals were obtained by
slow evaporation of a chloroform solution of iPr,PhLK. IR
(cm-1): ν(Ph2P-S) 586s, ν(iPr2P-S) 519s, ν(PNP) 1205br; 31P
NMR (121.75 MHz, CDCl3, 303 K): δP (ppm) = 36.8 [s, Ph2P],
72.4 [s, iPr2P].
Synthesis of Ni[iPr2P(S)NP(S)Ph2]2 (3). In a Schlenk-type
flask, 0.21 g (0.5 mmol) of iPr,PhKL were dissolved into 20 mL of
CH3OH. The mixture was stirred at room temperature for
30 min, until complete dissolution was reached and, subsequently,
0.059 g (0.25 mmol) of NiCl2 6H2O were added. The solution
3
turned gradually to dark green. After 2 h of stirring at room
temperature, a green solid precipitated that was filtered off and
washed with cold MeOH to yield 159.7 mg of 3 (78%). Dark-
green needle-like crystals, suitable for X-ray diffraction studies,
were grown by slow diffusion of n-hexane into a CH2Cl2 solu-
tion of 3 at 250 K. IR (cm-1): ν(Ph2P-S) 573s, ν(iPr2P-S)
527s, ν(PNP) 1231br; 1H NMR (300 MHz, CDCl3, 303 K):
The structural and refinement parameters for the iPr,PhLK ligand
and complex 3 are given in the Supporting Information, Table S1.
δ
H (ppm) = 7.8 [s, 4H, CH], 6.5 [br, 8H, m-C6H5], 6.03 [d, 4H,
5J(PH) = 58.18 Hz, p-C6H5], 7.77 [d, 12H, 3J(PH) = 123.49 Hz,
CH3], 8.41 [d, 8H, J(PH) = 141.25 Hz, o-C6H5], 9.45 [s(br),
3
12H, CH3]; 31P NMR (121.75 MHz, CDCl3, 303 K): δP (ppm) =
-512.4 [s(br), Ph2P], -555. [s(br), iPr2P].
(35) Wang, F. T.; Najdzionek, J.; Neneker, K. L.; Wasserman, H.;
Braitsch, D. M. Synth. React. Inorg. Met.-Org. Chem. 1978, 8, 119–125.
(36) Birdsall, D. J.; Slawin, A. M. Z.; Woollins, J. D. Polyhedron 2001, 20,
125–134.
(40) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098–3100.
(41) Perdew, J. P. Phys. Rev. B: Solid State 1986, 33, 8822–8824.
(42) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(43) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Solid State 1988, 37,
785–789.
(44) Miralles, J.; Castell, O.; Caballol, R.; Malrieu, J. P. Chem. Phys.
1993, 172, 33–43.
(45) Neese, F. J. Chem. Phys. 2003, 119, 9428–9443.
(37) Otwinowski, Z.; Minor, W. Methods in Enzymology, Macromolecular
Crystallography, Part A; edited by Carter, C. W., Jr, Sweet, R. M., Eds.;
Academic Press: New York, 1997; Vol. 276, pp 307-326.
€
€
(38) Sheldrick, G. M. SHELX97, University of Gottingen: Gottingen,
Germany, 1997.
(39) Neese, F. ORCA, version 2.6.35; University of Bonn: Bonn, Germany,
2008.

5082 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
π-electron density and thus its partial dissipation from the
P atom bearing the electron-donating iPr groups toward
the P atom bearing the electron-withdrawing Ph groups.
This electron density flow from “the alkyl towards the
phenyl side” is demonstrated by the different P-N bond
lengths and is also verified via 31P NMR spectroscopy
(vide infra).
On the contrary, the six-membered K-S-P-N-P-S
ring of the symmetric Ph,PhLK ligand adopts a perfect
twisted boat conformation, where both the P-N-P-S
torsion angles are equal (39.0ꢀ) and the Ph2P-S bonds are
orientated beneath and above the P-N-P plane, respec-
tively. The overall symmetry of Ph,PhLK is C2, with the
main axis defined by the K and N atoms and bisect-
ing both S-K-S and P-N-P angles.48 Therefore, the
different electron-donating properties of the periphe-
ral groups in iPr,PhLK distorts the overall symmetry and
gives rise to the boat conformation. It should be pointed
out that the conformation of the metallacycle is not
influenced by intermolecular interactions, since both
iPr,PhLK and Ph,PhLK exhibit the same polymeric structure.
Complex 3 crystallizes in the space group P-1 as two
crystallographically independent molecules 3a and 3b,
selected bond lengths and angles of which are listed in
the Supporting Information, Table S2. The only pro-
found structural difference between 3a and 3b is the
relative position of the methyl groups, i.e., the rotation
around the P-C bonds. The presence of two H atoms
near the vacant axial positions in 3b—their distance from
Figure 1. Thermal ellipsoid plot (50% probability) of the X-ray struc-
ture of the iPr,PhLK ligand and the centrosymmetric complex 3b.
Results and Discussion
Molecular Structures. The direct comparison between
the protonated and anionic forms of the dithioimidodi-
phosphinato ligands was up to now possible only for the
symmetric phenyl analogue, Ph,PhLH, given that the crys-
tal structure of both protonated46,47 and deprotonated
Ph,PhLK48 forms has been resolved. Herein, we report the
first crystal structure of a deprotonated asymmetric
ligand that belongs to the same family. iPr,PhLK crystal-
lizes in the space group P21/c. The ORTEP plot of
iPr,PhLK is shown in Figure 1, whereas selected bond
lengths and angles are presented in the Supporting Infor-
mation, Table S2. The P-N bonds are lengthened, the P-S
bonds are shortened, and the P-N-P angle is slightly
increased compared to the respective protonated iPr,PhLH
form.36 This is a well documented trend,21 originating
from the enhanced delocalization among the S-P-N-
P-S fragment and the ensuing higher sp2 character of the
nitrogen atom, upon deprotonation.33
˚
the central Ni atom being only 2.607 A—possibly stabi-
lizes the SP geometry. On the contrary, this close-contact
˚
interaction is absent (Ni H = 2.722 A) in the crystal
3 3 3
structure of 3a that exhibits a small tetrahedral distortion.
Regardless of whether this hypothesis is correct or not,
the tetrahedral distortion of 3a indicates that the stabili-
zation of the SP geometry for this family of NiS4 com-
plexes is marginal.
The ORTEP plot of the crystal structure of 3b is shown
in Figure 1. Complex 3b is centrosymmetric, while com-
plex 3a deviates from Ci symmetry, since both trans
S-Ni-S angles are less than 180ꢀ. The centrosymmetric
structure contains a planar NiS4 core (sum of S-Ni-S
bond angles 360ꢀ), and the two asymmetric iPr,PhL-
ligands are coordinated in a trans arrangement to each
other. The Ni-S/Ph bonds (see Supporting Information,
i
˚
Table S2) are slightly longer (2.236 A) than the Ni-S/ Pr
˚
bonds (2.223 A). Both bond lengths lie at the upper
The conformation of the M-S-P-N-P-S metalla-
cycles is governed by the P-N-P-S torsion angles that
define the orientation of the P-S bonds relatively to the
P-N-P plane. In the crystal structure of iPr,PhLK, the
Ph2P-S bond is rotated out of the PNP plane (P-N-
extreme of the range observed in SP NiS4 complexes
˚
9,11,49
(2.11-2.25 A)
and are significantly shorter than
0
the Ni-S bond lengths in Td NiR,R L2 analogues
11,23-26
˚
(2.28-2.30 A).
The comparison between the two
stereoisomers of the analogous Ni[iPr2P(Se)NP(Se)iPr2]2
i
P-S/Ph = 64.7ꢀ), whereas the Pr2P-S bond lies in the
complex supports the latter observation, since the Ni-Se
bond lengths were found shorter (2.354 A) for the SP than
P-N-P plane (P-N-P-S/iPr = 3ꢀ). Therefore, the
K-S-P-N-P-S ring exhibits a distorted boat confor-
mation, with P/Ph and S/iPr occupying the apexes. The
˚
32
˚
the Td isomer (2.400 A).
As it was discussed above, the extensive delocalization
of π-electronic density due to deprotonation and coordi-
nation of the ligand results in the lengthening of the P-S
bonds, the shortening of the P-N bonds, and the opening
of the P-N-P angle, as compared to the free protonated
i
fact that the Pr2P-S bond is almost coplanar with
the P-N-P backbone promotes the delocalization of
(46) Noth, H. Z. Naturforsch. 1982, 37b, 1491–1498.
(47) Husebye, S.; Maartmannmoe, K. Acta Chem. Scand. 1983, A37, 439–
441.
(48) Slawin, A. M. Z.; Ward, J.; Williams, D. J.; Woollins, J. D. J. Chem.
Soc. Chem. Commun. 1994, 421–422.
(49) Colpas, G. J.; Kumar, M.; Day, R. O.; Maroney, M. J. Inorg. Chem.
1990, 29, 4779–4788.

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5083
form. The comparison of the crystal structures of
iPr,PhLH,36 iPr,PhLK, and complex 3 can separate out the
geometric changes caused by either the deprotonation of
the ligand or its coordination to the Ni(II) center. As it can
be deduced from the selected bond lengths and angles
presented in the Supporting Information, Table S2, the
P-S bonds are further lengthened upon coordination,
while the P-N bonds remain practically unaffected.
Moreover, in complex 3, the boat conformation of the
Ni-S-P-N-P-S metallacycle is retained, with the P/Ph
and S/iPr atoms lying at the apexes. The iPr2P-S bonds
are located in the P-N-P plane (torsion angle
P-N-P-S/iPr = 0ꢀ), whereas the Ph2P-S bonds are
rotated out of it by 49ꢀ. The smaller torsion angle,
compared to iPr,PhLK, is a straightforward effect of the
shorter covalent Ni-S bonds compared to the longer
complexes (M = Fe,15,51-55 Co,15,16,51,52,55-59 Ni12-18) as
well as the analogous MiPr,iPrL2 complexes (M = Mn,33
Co,60 Zn25), which are always compressed along the
principal S4 symmetry axis.
The Td structure of 1Td only approximates a com-
pressed S4 symmetry, since the Ni-S bond lengths
and the S-Ni-S-P torsion angles are not exactly
equal,9,24 whereas the NiMe,MeL2 complex exhibits a
severely distorted Td geometry, of an approximate C2v
symmetry.23 On the other hand, when the ligands are
0
asymmetric (R ¼ R0), the corresponding NiR,R L2 com-
plexes display a SP geometry, with the Ni atom lying on a
center of inversion. As mentioned above, NiPh,PhL2 has
been isolated in both geometries (1Td and 1SP), being the
only pair of NiS4-containing stereoisomers reported up to
date. It should be stressed, however, that the crystal of 1SP
contains also solvent molecules. The SP complexes 1SP
and 3 revert to their Td paramagnetic isomers when they
dissolve in common organic solvents, as confirmed by
UV-vis, NMR, and magnetization studies, whereas
complex 1SP is transformed to 1Td even in the solid state
(vide infra). Only NiMe,PhL keeps its SP diamagnetic
geometry in solution.9 The fact that NiMe,MeL2 adopts a
distorted Td geometry23 suggests that the repulsion be-
tween the R peripheral groups is not the dominant factor
that promotes the Td geometry.
ionic K-S bonds. Therefore, the S S bite is decreased
3 3 3
iPr,Ph
˚
from 4.15 A in
˚
LK to 3.37 A in 3b, resulting in an
increased repulsive interaction between the S atoms and a
greater degree of ring strain among the ligands. This is
compensated, to a certain extent, by the relaxation of the
endocyclic S-Ni-S angles from the ideal value of 90ꢀ for
the ideal SP geometry to 98.2ꢀ in 3b. A similar value
has been reported for the Ni(II)-dithioacetylacetonato,
[Ni(SacSac)2], complex (97.2ꢀ).50 Moreover, in order to
reduce the ring strain and to accommodate the smaller
bite, the P-N-P angles are contracted from 131.1ꢀ in
iPr,PhLK to 124.2ꢀ in 3b. The Ph2P-S bonds move toward
the S-Ni-S plane (S-Ni-S-PPh2 = 15.4ꢀ), giving rise
to an overall ladder structure, in which the Ni-S-P-
The S-Ni-S endocyclic angles (see Supporting Infor-
0
mation, Table S3), of all the NiR,R L2 complexes under
investigation, can be compared with the ideal value of the
Td or SP geometry (109.5ꢀ and 90.0ꢀ, respectively). The
deviation is only 1-3ꢀ for the Td complexes, whereas it
exceeds 8ꢀ for the SP complexes. This is not an unusual
value for SP complexes containing six-membered chelat-
ing rings and has been ascribed to ring strain and poten-
tial repulsive interactions between the sulfur ligand
atoms.10,29 In contrast to the rest of the literature SP
NiS4 complexes,9 which usually contain a rigid planar
conjugated backbone and thus cannot tolerate very large
bite angles, the structural flexibility of the dithioimi-
dodip0hosphinate ligands allows the corresponding SP
NiR,R L2 complexes to revert to the less strained Td
geometry. Similar observations were made in the case of
the SP and Td isomers of the Ni[iPr2P(Se)NP(Se)iPr2]2
complex.32 We can, therefore, conclude that, although
this family of ligands can accommodate both geometries
upon coordination to Ni(II), the Td geometry is more
favorable, in contrast to what is common for Ni(II)
complexes bearing different S,S0-chelating ligands, which
usually exhibit SP geometry.9
N-S-P are folded by 51.3ꢀ toward the (iPr2P)S P(Ph)
3 3 3
direction.
0
Structural Preferences. Seven NiR,R L2 complexes have
been structurally characterized up to now, namely
NiPh,PhL2 (1Td),9,24 1SP
,
NiiPr,iPrL2 (2),25 NiMe,MeL2,23
28
NinBu,tBuL2,26 NiMe,PhL2,9 and NiiPr,PhL2 (3). The avail-
able crystallographic data show that when the ligands are
symmetric (R = R0), the corresponding complexes dis-
play a distorted Td geometry. Due to its high symmetry,
the structure of 2 would 0 serve as a reference structure
among the family of NiR,R L2 complexes. In this structure,
the N atoms are located in the S-Ni-S planes, while the P
atoms are rotated out of it toward opposite sides, with the
corresponding S-Ni-S-P torsion angles being exactly
equal (15.4ꢀ). Therefore, the Ni-S-P-N-P-S metalla-
cycle adopts a twist conformation, and the metal core of 2
displays an ideal S4 point group symmetry, with the main
axis bisecting both endocyclic S-Ni-S angles.25 Further-
more, the endocyclic S-Ni-S angles are slightly less than
109.5ꢀ (see Supporting Information, Table S3), thus the
core of 2 exhibits a minor tetragonal elongation, in con-
trast to what has been found for most Td [M(SAr)4]2-
Spectroscopic Studies. UV-vis Spectroscopy. The
electronic spectrum of 3, along with that of 1Td, in THF
is shown in the Supporting Information, Figure S1. The
spectrum of 3 shows a veryintense CT band at 26 180 cm-1
(50) Beckett, R.; Hoskins, B. F. J. Chem. Soc., Dalton Trans. 1974, 622–
625.
(51) Ueyama, N.; Okamura, T. A.; Nakamura, A. J. Chem. Soc., Dalton
Trans. 1992, 1019–1020.
(52) Okamura, T.; Takamizawa, S.; Ueyama, N.; Nakamura, A. Inorg.
Chem. 1998, 37, 18–28.
(53) Coucouvanis, D.; Swenson, D.; Baenziger, N. C.; Holah, D. G.;
Kostikas, A.; Simopoulos, A.; Petrouleas, V. J. Am. Chem. Soc. 1976, 98,
5721–5723.
(55) Huang, J.; Dewan, J. C.; Walters, M. A. Inorg. Chim. Acta 1995, 228,
199–206.
(56) Fukui, K.; Masuda, H.; Ohyanishiguchi, H.; Kamada, H. Inorg.
Chim. Acta 1995, 238, 73–81.
(57) Otto, J.; Jolk, I.; Viland, T.; Wonnemann, R.; Krebs, B. Inorg. Chim.
Acta 1999, 285, 262–268.
(58) Walters, M. A.; Dewan, J. C.; Min, C.; Pinto, S. Inorg. Chem. 1991,
30, 2656–2662.
(54) Coucouvanis, D.; Swenson, D.; Baenziger, N. C.; Murphy, C.;
Holah, D. G.; Sfarnas, N.; Simopoulos, A.; Kostikas, A. J. Am. Chem.
Soc. 1981, 103, 3350–3362.
(59) Chung, W. P.; Dewan, J. C.; Walters, M. A. J. Am. Chem. Soc. 1991,
113, 525–530.
(60) Gilby, L. M.; Piggott, B. Polyhedron 1999, 18, 1077–1082.

5084 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
respect, band 6 corresponds to a “d-d” transition, strongly
overlapped by CT bands.62 If we assume a Td symmetry for
the NiS4 core, the low-energy band 2 (10 200 cm-1, ε = 110
M
3
-1cm-1) is assigned to the ν2 T1(F) f 3A2(F) transition,
whereas the bands 3-6 derive from transitions to the same
excited term split by tetragonal elongation and spin-orbit
coupling effects. This multicomponent band corresponds to
the ν3 (3T1(F)) f (3T1(P)) transition. By using the corre-
sponding Tanabe-Sugano diagram, estimates of Δt =
5300 and B0 = 812 cm-1 are derived. Broadening of the
peaks and distortion of the NiS4 core to approximately D2d
symmetry leads to some uncertainty in assigning the center
of the bands in terms of Td symmetry. However, these
values are consistent with those found for complex 1Td.27
Figure 2. The UV-vis spectrum of complex 2 in CH2Cl2 fitted with
Gaussian curves up to the “d-d” and CT transitions region. The inlet
highlights the vis/NIR region of the spectrum.
Since the value of B for Ni2þ(g) ion is 1080 cm-1 61
the
,
nephelauxetic parameter β of complex 2 is approximately
equal to 0.75, indicating a significant degree of covalency
in the Ni-S bonds of such Ni(II)S4 systems. In the theore-
tical section, a full assignment of the bands, provided
by multiconfiguration ab initio correlated methods, will
be presented.
IR Spectroscopy. IR spectroscopy has been proved to
be very helpful in investigating the effects exerted by the
0
different peripheral R groups in MR,R L2 complexes.21
The IR spectrum of iPr,PhLK is presented here for the first
time (Supporting Information, Table S4). Deprotonation
of the iPr,PhLH ligand leads to the loss of the v(N-H)
band. The v(P-N-P) band is shifted to higher frequen-
cies, while the v(P-S) bands are shifted to lower frequen-
cies, compared to iPr,PhLH,36 in complete agreement with
the shortening of the P-N bonds, the lengthening of the
P-S bonds, and the overall enhancement of delocaliza-
tion of π-electronic density upon deprotonation. A
similar trend has been documented for the IR spectra
Figure 3. The metal-d-based MO and term symbols (analyzed under
approximate D2d symmetry), arising from single excitations in the elon-
gated Td NiH,HL2 model. The indicated orbital occupation pattern refers
to the 3A2 ground state.
of Ph,PhLK,21,48 Me,PhLK9 and for all the homoleptic
0
M
R,R L2 complexes.21 The above effects are manifested
also in the IR spectrum of 3 (Supporting Information,
Table S4). Moreover, it is noteworthy that the ν(P-N-P)
bands of all the SP complexes are ∼50 cm-1 lower than
the corresponding bands of the Td complexes, a fact that
is attributed not only to weaker P-N bonds but also to
smaller P-N-P angles, since the “double-bond charac-
ter” of the P-N-P fragment is decreased. This is corro-
borated by the crystal structures of the corresponding
complexes, in which the P-N-P angles of the Td species
are significantly larger (by 10ꢀ or more) compared to the
SP species (Supporting Information, Table S3). In that
respect, the comparison between the 1Td (P-N-P =
130ꢀ) and 1SP complexes (P-N-P = 120ꢀ) is rather
characteristic.
(ε = 5350 M-1cm-1) and that of 1Td at 25 870 cm-1 (ε =
7060 M-1cm-1), with a shoulder at 20 980 cm-1 (ε = 1010
M
-1cm-1) and a low intensity feature which consists of
two broad bands at 14 110 (ε = 154 M-1cm-1) and 12 760
cm-1 (ε = 175 M-1cm-1). The latter is assigned to the ν3
3T1(F) f 3T1(P) transition and is characteristic of Td
complexes with a distorted NiS4 core.27,61 The close resem-
blance of the two spectra suggests that, upon dissolution,
complex 3does not retain the SP geometry established in the
solid state by X-ray diffraction, but it converts to a Td NiIIS4
chromophore. This is further supported by 31P NMR
spectroscopy (vide infra). In Figure 2, a combined near
IR-UV/vis absorption spectrum is presented for complex 2
in CH2Cl2. The Gaussian resolution obtained by simulta-
neous fitting of the spectrum gives rise to five distinguish-
able absorption bands for the lower-energy section of the
spectrum, which reflects a distorted Td Ni(II)S4 core.
More specifically, the 1-6 bands correspond to transitions
of “d-d” character, bearing in mind that the highest
occupied-singly occupied molecular orbitals (HOMO-
SOMO) involved are molecular orbitals with a significant
ligand contribution, due to the covalent character of the
Ni-S bonds (vide infra), as illustrated in Figure 3. In that
NMR Spectroscopy. The 31P-{1H} NMR spectrum of
3 (Supporting Information, Figure S3) shows two sepa-
rate peaks, assigned to the two nonequivalent P atoms of
the ligand, without any well-resolved hyperfine cou-
plings. The NMR signals lie in the far paramagnetic
region of the spectrum, supporting the conversion of 3
to a Td NiS4-containing site in solution, in agreement with
the UV-vis data described above as well as the magnetic
measurements (vide infra). It should be stressed that a
similar SP f Td isomerization takes place between 1SP
(61) Figgis, B. N.; Hitchman, M. A. Ligand Field Theory and Its Applica-
tions; Wiley-VCH: New York, 2000.
(62) Nieto, I.; Bontchev, R. P.; Ozarowski, A.; Smirnov, D.; Krzystek, J.;
Telser, J.; Smith, J. M. Inorg. Chim. Acta 2009, 362, 4449–4460.

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5085
0
0
Table 1. 31P-{1H}-NMR Spectroscopy Data of R,R LK Ligands and MR,R L2
flow is always present, regardless of the overall geometry.
However, the diamagnetic nature of SP complexes limits
this electron density flow.
Complexesa
31P chemical shifts (ppm)
The latter seems to play a substantial role in the stabi-
lization of the boat conformation of the Ni-S-P-
N-P-S chelating rings. In all the aforementioned com-
plexes, the P atom that occupies one apex of the boat
(hence the corresponding P-S bond is located out of the
P-N-P and in the S-M-S plane) always “carries” Ph or
other even stronger electron-withdrawing groups, like
OPh and OEt. However, the stabilization of the boat
conformation cannot be simply explained on the basis of
different electron-donating or -withdrawing effects. In
the crystal structure of PdEtO,PhL2, it is the Ph2P-S bond
that lies in the coordination plane (i.e., PPh2 at the
apex), although OEt is considered as a stronger electron-
withdrawing group.67 The tendency of the Ph2P-S bonds
to lie in the main coordination plane does not only stem
from the electron-withdrawing properties of the Ph
groups but also from their “in plane” orientation as well,
which enhances the delocalization of electron density on
this highly delocalized system.
Magnetic Properties of the SP Complexes. In order to
study possible interconversions between the SP and Td
species, magnetic susceptibility measurements of the SP
complexes 1SP and 3 were conducted. Negligible evidence
of any paramagnetic behavior has already been found for
NiMe,PhL2, since it maintains its diamagnetic SP geometry
both in the solid state and in solution.9
The results of the magnetic susceptibility measure-
ments show that for complex 1SP, only 4.8% of the sample
exists as a paramagnetic Td compound at low tempera-
tures (<40 K). However, complex 1SP reverts back to its
Td isomer (1Td) at room temperature over time, indicat-
ing that the SP geometry is less stable than the Td one.
Measurement of the magnetic susceptibility for 3 reveals a
similar, mainly diamagnetic signal with a small Curie tail,
compound
Ph2P
38.0
—
36.8
-32.8
—
-87.8
-437.2
—
R2P
2J (P-P)
Ph,PhLK48
—
60.8
72.4
iPr,iPrLK (this work)
iPr,PhLK
13.1
CoPh,PhL2 33,60
60
—
CoiPr,iPrL2
-331.9
-188.4
—
33
CoiPr,PhL2
Td NiPh,PhL2 (1Td
NiiPr,iPrL2 (2)25
NiiPr,PhL26(33)
PdPh,PhL2 64
)
—
—
9,24
-725 (br)
-553 (br)
—
-509.9
38.7
PdiPr,iPrL236
64.8
69.3
—
62.1
74.1
PdiPr,PhL2
37.1
45.1
46.1
45.2
12.8 Hz
—
15.3 Hz
9.8 Hz
PdPhO,PhOL2
65
PdPhO,EtL2 36
36
PdPhO,iPrL2
a Solvent: CDCl3, 298 K.
28
and 1Td as well as between the SP and Td isomers of
Ni[iPr2P(Se)NP(Se)iPr2]2.32 On the contrary, based on
UV-vis and NMR data, the SP NiMe,PhL2 complex seems
to retain its diamagnetic nature in solution.9
The 31P chemical shifts of complexes 1-3 are collected
in Table 1. From the comparison between the chemical
shifts (Supporting Information, Figure S3), it becomes
clear that in 3, which contains the asymmetric
[iPr2P(S)NP(S)Ph2]- ligand, the PPh2 atom is shielded,
whereas the PiPr2 atom is deshielded, compared to the
corresponding symmetric0 complexes 1Td and 2. For the
analogous series of CoR,R L2 complexes, we have already
proposed the existence of an electron density flow from
i
the P atom bearing the electron-donating Pr groups to
the P atom bearing the electron-withdrawing Ph groups,
via the central metal atom.33 The spectroscopic study of
0
complexes 2 and 3 completes the NiR,R L2 series and
further supports our proposal. The stronger paramagne-
tic shifts observed for the Ni(II) complexes (δ = -437 to
-725 ppm) compared to the analogous Co(II) complexes
(δ = -33 to -188 ppm)33 is indicative of greater spin
density on the ligands for the Ni(II) systems, consistent
with the magnetic data analysis (vide infra).
giving a 3% paramagnetic impurity. In contrast to 1SP
,
there is no rise of the magnetic susceptibility of 3 over time
and temperature, implying that, in the solid state, 3
preserves its SP geometry. However, a paramagnetic
nature of 3 is inferred in solution by UV-vis and NMR
studies (vide supra). It can thus be concluded that the
geometric preferences of 1SP, 3, and NiMe,PhL2 are finely
balanced. It appears that complex 1 prefers a SP geometry
only at very low temperatures, with specific solvents of
crystallization also being important.28 Complex 3 is SP in
the solid state even at room temperature, but it adopts a
paramagnetic Td geometry in solution. Finally, NiMe,PhL2
is the most stable SP complex among the three, since it
preserves its geometry in solution.9
We cannot make a similar comparison for the SP Ni(II)
complexes, given the fact that 3 and 1SP revert to their Td
paramagnetic isomers upon dissolution. However, a clo-
0
ser examination of the corresponding PdR,R L2 com-
plexes,36,63-66 which retain their SP geometry in solution,
reveals that the P atoms bearing alkyl groups are de-
shielded, whereas the P atoms bearing the electron-with-
drawing Ph or OPh groups are shielded, again compared
to the corresponding symmetric complexes. Moreover,
this trend is apparent in the 31P NMR spectra of the
potassium salts of the ligands (Table 1). Consequently, we
can postulate that the aforementioned electron density
(67) Cupertino, D. C.; Keyte, R. W.; Slawin, A. M. Z.; Woollins, J. D.
Polyhedron 1999, 18, 311–319.
(68) Figgis, B. N.; Lewis, J.; Mabbs, F. E.; Webb, G. A. J. Chem. Soc. A
1966, 1411–1421.
(69) Boca, R. Coord. Chem. Rev. 2004, 248, 757–815.
(70) Boca, R. Structure and Bonding. Magnetic Functions beyond the
Spin-Hamiltonian; Mingos, D. M. P., Ed.;Springer-Verlag: Heidelberg,
Germany, 2006; Vol. 117, pp 1-264.
(63) Abbati, G. L.; Aragoni, M. C.; Devillanova, F. A.; Fabretti, A. C.;
Garau, A.; Isaia, F.; Lippolis, V.; Verani, G. J. Chem. Soc., Dalton Trans.
2001, 1105–1110.
(64) Cupertino, D.; Keyte, R.; Slawin, A. M. Z.; Woollins, J. D.;
Williams, D. J. Polyhedron 1996, 15, 4441–4445.
(65) Nouaman, M.; Zak, Z.; Herrmann, E.; Navratil, O. Z. Anorg. Allg.
Chem. 1993, 619, 1147–1153.
(71) Maganas, D.; Kyritsis, P.; Groenen, E. J. J.; Neese, F. Manuscript in
preparation.
(72) Fukui, K.; Kojima, N.; Ohyanishiguchi, H.; Hirota, N. Inorg. Chem.
1992, 31, 1338–1344.
(66) Cupertino, D.; Birdsall, D. J.; Slawin, A. M. Z.; Woollins, J. D. Inorg.
Chim. Acta 1999, 290, 1–7.
(73) Vrajmasu, V. V.; Munck, E.; Bominaar, E. L. Inorg. Chem. 2004, 43,
4867–4879.

5086 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
Figure 4. Magnetic susceptibility against temperature for (a) 1Td and (b) 2, parallel and perpendicular to the unique axis. The red continuous lines were
simulated using the procedure described by Figgis et al.68 (c) The blue lines are fits to the experimental curves of complex 2, with the close form of magnetic
susceptibility equations, using the calculated g^, g// and D parameters, as discussed in the Magnetic Properties Calculation Section (vide infra).
Table 2. Data Derived from the Analysis of the Magnetic Properties of
Complexes 1Td and 2
employed by Figgis et al.68 However, we believe that the
data obtained for complex 1Td, albeit less accurate,
complement those obtained for 2, reinforcing the conclu-
sions drawn.
axial field
component,
spin-orbit
coupling
orbital
reduction
factor, k
Δ (cm-1
)
constant, λ (cm-1
)
The cubic crystal field parameter was found to corres-
pond to a weak field in both cases, as has been ob-
served previously for Td complexes which are typically
weak field systems. This is also consistent with the
UV-vis analysis (vide supra). The axial component
of the crystal field was found to be relatively small for
both complexes, and this is consistent with the small
deviations from Td geometry observed in the structural
data. The most interesting parameters in this system
are the orbital reduction factor (k) and the spin-orbit
coupling constant (λ). The orbital reduction factor
gives a measure of the spin density located on the metal
center and varies between 1 and 0, corresponding
to all the spin density located on the metal or all on the
ligands, respectively. It should be noted that the values
of k and λ should be considered together when interpret-
ing physical properties of the complexes, as both para-
meters give an indication of spin density delocalization
onto the ligands. Following the method by Figgis et al.,68
k is introduced as a coefficient of λ when the Zeeman
perturbation is applied, after the spin-orbit coupling
perturbation. In interpreting the physical meaning of
these two parameters, based on the low values of
both, the conclusions described below are drawn but
care is needed not to overinterpret the individual mag-
nitude of each.
1Td
2
110
200
-45
-79
0.5
0.35
Magnetic Properties of the Td Complexes. The tetra-
hedral complex 1Td packs in a monoclinic unit cell that
contains two independent molecules.9,24 However, the
b and c axes are almost equal and perpendicular to each
other. Moreover, the molecules are oriented with the long
molecular axis along the a direction. Thus, the two
different molecules and the b and c axes have been
regarded as equivalent, in order to simplify the magnetic
analysis. The Td complex 2 packs in a tetragonal unit cell,
where the a and b axes are equivalent, and the c axis is
unique. Variable-temperature magnetic susceptibility
measurements were carried out on powder and single
crystal samples of 1Td and 2, with the applied magnetic
field parallel and perpendicular to the unique z axis,
which is defined as the N-Ni-N direction. Following
the method described by Figgis et al.,68 we applied the
magnetic field perturbation (parallel and perpendicular
to the unique axis) to the wave functions of the 3T1 term,
split by spin-orbit coupling and axial ligand field, and
obtained the first- and second-order coefficients for sub-
stitution into the Van Vleck equation that describes the
magnetic moment as a function of temperature. Due to
the complexity of the resulting expression, we used trial
values of the crystal field strength, the spin-orbit cou-
pling constant (λ), the axial crystal field component (Δ),
and the orbital reduction factor (k) to simulate the
temperature-dependent susceptibility curves. The input
parameters were altered by inspection, until an acceptable
simulation of the experimental data was achieved (Figure
4a and b). A wide range of parameter combinations was
investigated, and those reported are the only combina-
tions that satisfactorily simulate the experimental data.
Small discrepancies may be attributed to deviations from
Simple inspection of the magnetic susceptibility curves
for 1Td and 2 (Figure 4a and b) indicates that these
complexes differ greatly from literature examples of Td
Ni(II) complexes,68 since the peak in χz and the plateau in
χx occur at much lower temperatures than is typical. The
origin of this effect is revealed by the extremely low values
of both k and λ for 1Td and 2 (Table 2).
In comparison, for a typical series of related
complexes,68 k takes values from 0.55 to 1.00, and λ varies
between -130 to -224 cm-1, with the free ion value
at -315 cm .
-1 61 The very small orbital reduction factor
ideal axially distorted tetrahedral geometry in the crystal,
which are larger for complex 1Td
gives a direct indication of the reduced spin density on the
metal and, furthermore, the low value of the spin-orbit
coupling constant arises from the same characteristic of
the system. We are unaware of any prior examples where
values of k and λ are both so low and interpret this unique
behavior as arising from the noninnocent characteristics
of the ligand, leading to a large amount of spin density
located on the ligands.
9,24
compared to 225 as
well as to imperfect alignment of the crystals parallel and
perpendicular to the field direction. The resulting para-
meters derived for 1Td and 2 are given in Table 2. The data
obtained for complex 2 are considered as more accurate,
since its structural characteristics are closer to an axially
distorted tetrahedral field, as required by the analysis

Article
In addition to the above parameters, for complex 2,
Inorganic Chemistry, Vol. 49, No. 11, 2010 5087
Scheme 1a
which exhibits an axially elongated S4 geometry, it is
meaningful to extract from our simulation the splitting
between the ms = 0 ground state and the lowest-lying
ms = ( 1 excited states,69,70 as described explicitly in the
Supporting Information. This corresponds to the axial
component, D, of the zero-field splitting (ZFS) employed
in the spin Hamiltonian (SH) description of the S = 1
system (Supporting Information, Figure S2). A value of
77 cm-1 is estimated for complex 2, which compares well
with the calculated ZFS D values based on the theoretical
calculations (vide infra). The analysis described above,
comprising determination of the first- and second-order
Van Vleck coefficients, considers the action of the spin
and orbital components of the magnetic moment opera-
tor on the wave functions derived from the 3T1 state. It is
limited by the need to consider an axial system, whereas
the SH approach, which does not explicitly consider
orbital angular momentum, enables some rhombic char-
acter to be considered through the parameter E (vide
infra). It is important to note that the resulting calculated
value of E is negligibly small and that the values of D,
determined by both approaches, are very similar. The
consistency between the two approaches suggests that both
are valid in providing a good description of the system.
Finally, in order to put the structural, magnetic, and
electronic properties of complexes 1Td and 2 into perspec-
tive, we should make clear that they exhibit subtle struc-
tural differences, since they possess a tetragonally
compressed and elongated NiS4 core, respectively. This
is expected to lead to different ground and excited states,
as explicitly described when the effects of the R peripheral
groups on the electronic properties of the system are
taken into account (vide infra).
a The two E-symmetry bending modes Qθ (left) and Qε (right), which
dominate the Jahn-Teller (JT) distortion in Td complexes.
(90ꢀf0ꢀ) or the pairs of twisting exo-S-Ni-S angles R2
(109.5ꢀf180ꢀ).
Distortions along the two components of the E sym-
metry bending modes lead to two different SP conforma-
tions that are, in an idealized case, of D4h and D2h
symmetry. For bidentate ligands, Qθ cannot be operative
because of ring strain. Therefore, it is mainly Qε that
dominates the interconversion pathway. The Td f SP
interconversion pathway for complexes containing
chelating ligands has been shown to follow the sequence
D2d f D2 f D2h.74 For making comparisons easier, 0we
will treat the isomerization procedure in the NiR,R L2
complexes by keeping the same symmetry notation.
In principle, there are two possible electronic config-
urations for a d8 tetragonally distorted Td system under
2 2
the D2d point group notation: (i) [b2 e ] for δ < δt, and
1 3
(ii) [b2 e ] for δ > δt . These are related via two single elec-
tron excitations, namely dxy f dxz or dxy f dyz (Figure 3).
The D2d f D2h isomerization proceeds via known routes,
depending on the ground-state configuration of the D2d
structure and the number of Q modes that dominate.74 If
the twisting force along Qε is negligible, the value of δ will
determine the ground-state configuration.75 However,
when both forces are active, the value of δ cannot provide
a safe prediction of the triplet ground-state electron
Quantum Chemical Calculations. Geometric Consid-
erations. Ni(II) (3d8) may form Td, paramagnetic (high-
spin, triplet, S = 1) or SP, diamagnetic (low-spin, singlet,
S = 0) complexes, which exhibit very different properties.
As discussed above, the Td NiS4 complexes display S4
configuration.
0
The Td f SP interconversion pathway in the NiR,R L2
complexes will be discussed in terms of: (i) spin-orbit
coupling (SOC) in the ground state and (ii) Jahn-Teller
(JT)-active excited states. The significant mixing of states
provided by strong SOC interactions triggers and, in
some cases, completely quenches the JT distortion forces.
Correlation between Ring Conformation and Geometry
in the SP Complexes. The M-E-P-N-P-E metalla-
cycles in transition-metal complexes of the general type
M[R02P(E)NP(E)R2]2 (E = O, S, Se) can adopt different
conformations (Supporting Information, Figure S4).
Most of the Td complexes of the first-row transition
metals exhibit the same twist conformation, as the one
observed in the crystal structure of complex 2 (vide
supra). On the contrary, in all the SP complexes of
elongated or compressed structures. However, in the
0
corresponding CoR,R L2 complexes, it has been shown
that the EPR properties, investigated by both experi-
mental34 and theoretical71 methods, match those of [Co-
(SAr)4]2- systems of D2d symmetry.72 In particular, using
the annotation defined for the FeS4-containing proteins
rubredoxins, this structure approximates D2d(1),73 being
only lower due to ligand-constraining forces (i.e., the
six-membered ring is not planar, torsion angle S-Ni-
S-P > 0). Thus, for convenience we will adopt the sym-
metry labeling of the D2d point group.
Interconversion Pathway between SP and Td Geome-
tries. Qualitative Discussion. It has been recognized that
the reaction coordinate corresponding to the interconver-
sion between the triplet Td and the singlet SP states starts
along a nuclear displacement that belongs to the repre-
sentation E of the Td symmetry point group. The two
degenerate modes of E symmetry are shown in Scheme 1.
Qθ represents a nuclear motion which would result in
compressing the ligands into the xy plane, while Qε is a
twist, tending to put the ligands into the xz plane. Qθ is
defined by the δS-Ni-S angle along the S4 principal axis
(for an ideal Td geometry δt = 109.5ꢀ), while Qε is defined
as the angle between the S-Ni-S/S-Ni-S planes R1
Ni(II) 9,28
the sulfido analogues, the metallacycles adopt the boat
conformation (boat-SP) while, in the case of Pt
(II) 19 the
,
including complex 3, and Pd(II)36,63-65 bearing
,
chair conformation (chair-SP). The twist conformation
in a SP complex is only encountered in the M[iPr2P-
(Se)NP(Se)iPr2]2 complexes, where M = Ni,32 and Pt.66
(74) Lohr, L. L., Jr; Grimmelmann, E. K. J. Am. Chem. Soc. 1978, 100,
1100–1105.
ꢀ
(75) Mihail, A.; Cedrick, R.; Pio, B.; Claude, D. Int. J. Quantum Chem.
2005, 102, 119–131.

5088 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
Figure 5. Correlation between the conformation of the Ni-S-P-
N-P-S metallacycle and the S-Ni-S bite angle.
Figure 6. DFT BP86/TZVP constructed interconversion pathway along
the R2 exo-S-Ni-S twisting angle, byemploying relaxed surface scans for
the S = 1 and 0 states of NiH,HL2 model complexes. The vertical line
defines the R2 > R2t (right) and R2 < R2t (left) regions of the pathway.
Since the P atoms that lie on the same side of the
coordination plane are coordinated in a trans arrange-
ment, we call this conformation trans-twist-SP. This
results in an overall D2 symmetry and corresponds to the
D2h structure of the D2d f D2h interconversion pathway,
as found for the Ni(II) complexes containing chelating
ligands, being only lower due to ligand-constraining
forces.74,76 We have performed relaxed potential scans
along the S-Ni-S bite angle on the NiH,HL2 model,
keeping the R2 angle constrained to 180ꢀ. From the
diagram presented in Figure 5, it is evident that the
chair-SP conformation can better accommodate smaller
bite angles close to 90ꢀ, while, as the S-Ni-S angle opens
identified along the interconversion pathway (points
A-F; Figure 6 and Supporting Information Figure S4).
A complete descriprion of the DFT-constructed pathway
with respect to the structures identified along it is pro-
vided in the Supporting Information. It should be high-
lighted that, according to the DFT geometry optimiza-
tions, the S = 1 ground state is unstable around point A
of Figure 6, which is contradictory to the experimental
data. This inconsistency originates from a characteristic
issue of DFT which, as a single determinant method, may
fail to yield accurate electronic structures in which two or
more configurations mix. Under D2d symmetry, the two
3A2 states (originating from 3T1 in Td symmetry) mix and
reduce the JT stabilization energy (EJT).77 Despite this
shortcoming, DFT methods are still useful to extract
valuable information for the ground-state composi-
tion, along the interconversion pathway. Therefore,
the ground-state electronic configuration observed at
the starting point A of the interconversion pathway
up and the S S distance increases, the boat-SP con-
3 3 3
formation is promoted. The cis-twist-SP conformation
was considered as well but was found to be a transition
state that relaxes back to the boat-SP conformation. It is
concluded that, as the value of the bite angle increases, the
chelating ring conformation changes from chair (∼90ꢀ) to
boat (∼94ꢀ-98ꢀ) to twist (∼99ꢀ-104ꢀ). This is consistent
with the experimentally observed trend. Factors that
directly influence the E-M-E0 bite angle, such as the
nature of the metal center (Ni, Pd, Pt) or the E donor
atom (S, Se, Te),32 would lead to different ground-state
SP conformations.
2 2
(Figure 6) is b2 e , which corresponds to an elongated
Td structure along the principal S4 axis and is not JT in
nature. However, the singly excited states of 3E symmetry
1 3
Quantum Chemical Calculation of the Interconversion
Pathway. It is well-known that the Td f SP interconver-
sion pathway is dominated by the strengthening of
the ligand field (LF) as well as by orbital relaxation
effects.10,29 An additional insight into these phenomena
can be provided by the theory of the pseudo-Jahn-Teller
effect (PJTE).
(b2 e ) are subject to JT distortions, and they are expected
to destabilize the ground state via PJTE.78 In fact, by
more carefully mapping the twist R2 angle region from
90ꢀ-130ꢀ in steps of 1ꢀ, three well-distinguishable regions
€
emerge (Figure 7a). From a Lowdin population analysis,
itcan be seenthatinthe108ꢀ<R2 < 117ꢀ region, theS = 1
state corresponds to the 3A2(b2 e ) electron configuration
2 2
Initially, we have constructed the DFT BP86/TZVP
relaxed scanned potential energy surfaces for the triplet
and singlet states, along the exo-S-Ni-S twisting angle
R2. The scan along R2 > Rt dominates the D2d f D2 f Ci
(A-C-D) interconversion pathway, while the scan along
R2 < Rt dominates the D2d f C2v pathway, respectively,
in agreement with previous findings.10,29 In the particular
case of R2 > Rt, the MECP spin state occurs at R2 = 140ꢀ
(point C). This spin-crossover state lies only 3 kcal/mol
higher than the Td ground state, indicating a facile Td f
SP conversion. Overall six characteristic structures can be
and displays approximate S4 symmetry. Outside this
region (R2 < 107ꢀ or R2 > 118ꢀ), the dominant compo-
nent of the triplet state originates from the E(b2 e )
3
1 3
configurations (corresponding to dxy f dxz and dxy
f
dyz excitations, respectively). It is concluded that, as the
value of the δ angle deviates from δt, mixing of excited
states that are JT-active destabilizes the ground state.
Consequently, distorted Td structures are observed,
1 3
which are correlated with the E ꢀ e JTE of the 3E(b2 e )
(77) Reinen, D.; Atanasov, M.; Nikolov, G. S.; Steffens, F. Inorg. Chem.
1988, 27, 1678–1686.
(76) Starikov, A. G.; Minyaev, R. M.; Minkin, V. I. Chem. Phys. Lett.
2008, 459, 27–32.
(78) Garcia-Fernandez, P.; Bersuker, I. B.; Boggs, J. E. J. Chem. Phys.
2006, 125, article number: 104102.

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5089
Figure 8. (a) Stabilizationofthe PJTE. MRDDCI2(7,12) energiesofthe
3E(b21e3) and 1E(b22e2e0) states, respectively, inducing the spin crossover
at the twisting angle R2 = 161ꢀ.
1 3
other hand, the 3E(b2 e ) (dxy f dyz) state is stabilized for
R2 > 107ꢀ and dominates the ground spin state for R2 >
117ꢀ, showing a rather shallow minimum at R2 ∼ 120ꢀ
2
1
(dyz dxy dxz1 electron configuration).
Dynamic Correlation Contributions. More accurate
results can be obtained by including dynamic correlation
contributions on top of the SA-CASSCF wave func-
tions. To this end, MR-DDCI2(12,7) calculations have
been performed (Figure 8). From these calculations, it is
3
1 3
found that the triplet E (b2 e ) JT-active excited state
stabilizes for twisting angles R2 > 119ꢀ compared to
3
2 2
the A2(b2 e ) ground state, in agreement with the SA-
CASSCF calculations (Figure 7b). On the other hand, the
Figure 7. Identification of the twisting angle dependence of the S = 1
ground state, along the interconversion pathway derived from (a) DFT
and (b) SACASCF(12,7) calculations.
2 2 0
singlet 1E (b2 e e ) JT-active excited state dominates the
ground state for R2>161ꢀ. It should be noted here that
inclusion of dynamic correlation on top of the SA-
CASSCF(12, 7) energies proved critical in order to pre-
dict such singlet triplet energetic crossings. The JT stabi-
state. We have so far treated the important spin-conser-
ving excitations, which influence the ground state along
the interconversion pathway. However, the Td f SP
isomerization of a d8 transition-metal complex is a pro-
cess that involves a transition between two states of
different spin multiplicities. We, therefore, deal with
two cooperative phenomena: a PJTE which involves
spin-conserving excitations and a second procedure
which involves spin-flipping excitations. Although such
excitations are forbidden in high-spin, undistorted geo-
metries, they might become allowed in distorted high-spin
geometries due to enhanced molecular orbital mixing78,79
that further stabilizes the low-spin configuration.
3
1 3
lization energy for the E (b2 e ) (dxyfdyz) triplet state
is calculated to be EPJT1 = 21.2 kcal/mol at R2 = 120ꢀ
(Figure 8).
In agreement with previous findings,74,77 under strong
state-mixing conditions, the low-spin excited state of the
undistorted geometry becomes the lowest in energy in the
distorted configuration. Such distortion is accompanied
by d orbital mixing, which lowers the symmetry of the
charge distribution. However, as discussed above, a spin-
flip excitation is necessary for the stabilization of the low-
1
spin SP state, which can be thought as a (3A2þ E) ꢀ e
2 2 0
PJTE distortion of the b2 e e state, with dxy dxz dyz
2
2
0
A more detailed approach can be provided via wave
function-based multireference ab initio methods that treat
the interacting states in a simultaneous and balanced way.
State-averaged CASSCF(12,7) (SA-CASSCF) calcula-
tions were performed at the BP86 optimized structures
(Figure7b). Again, the energy of thethreedifferentelectron
configurations is plotted versus the twisting angle R2. The
failure of the DFT calculations to predict an energetic
minimum for the 3A2 ground state is alleviated in the SA-
CASSCF calculations. A distinct local minimum is found
electron configuration. This phenomenon, i.e., that an
overall ground-state structure originates from the distor-
tion of a low-spin structure of higher energy than its high-
spin counterpart, may be considered as a JT-induced spin
crossover.78,79 The MR-DDCI2 (12,7) JT stabilization
energy for the 1E singlet is EPJT2 = 28.9 kcal/mol (R2 =
1
175.2ꢀ). In fact, the E state is stabilized in close agree-
ment with the DFT results but sort of shifted (161ꢀ with
respect to 145ꢀ in DFT) and dominates the S = 0 ground
state for the twisting angles range of R2 = 160ꢀ-180ꢀ.
Most important is the observation that the energy mini-
mum of the 1E state is larger by 17.1 kcal/mol compared
to that of the corresponding 3A2 state, and in accordance
3
for twisting angles between 107ꢀ and 117ꢀ, with the A2-
(b2 e ) configuration clearly dominating the ground state.
2 2
1 3
On the other hand, the 3E(b2 e ) (dxy f dxz) stabilizes for
R2<107ꢀ and dominates the ground state for R2<103ꢀ. The
global stabilization minimum is at R2 ∼ 100ꢀ and corres-
2
1
ponds to the dxz dxy dyz1 electron configuration. On the
(79) Bersuker, I. B. Mol. Struct. 2007, 838, 44–52.

5090 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
the net result of such strong state mixing is the quenching
of the instability of the JT-active E state by lifting its
3
degeneracy, Supporting Information, Figure S5. The
SOC even at around the highest symmetry point δ ∼ δt
and R2 ∼ Rt is sufficiently large to change the adiabatic
potential energy surface (APES) morphology around it,
from the JT “Mexican hat” APES to a parable with
positive curvature (Figure 7b). We can, therefore, con-
clude that the ab initio treatment of the induced PJT
interconversion pathway provides a complete description
of the experimental structures observed, which otherwise
would have led to incomplete structural explanations.
R Peripheral Group Effects. So far we have studied the
electronic factors that promote the D2d f D2 f Ci path-
way. However, there are some drawbacks which do not
allow the treatment of the problem as solely induced by
JT phenomena. As discussed above, complexes with
symmetric ligands will tend to be stabilized at the S = 1
D2d structure, as is fact the case experimentally. In this
section, the stereochemical factors that favor or oppose
the action of the electronic factors will be included in the
discussion of the observed X-ray structures. Symmetric
NiR,RL2 complexes will be stabilized along the intecon-
version pathways, depending on their0 δ and R values.
Complex 2 is the only elongated MR,R L2 complex with
zero twisting force, in agreement with the DFT calcula-
tions of the D2d NiH,HL2 model structure. We can, there-
fore, conclude that this structure is not PJTE influenced,
Figure 9. The MO diagram for the D2d f C2(Ci) isomerization of the
0
NiR,R L2 complexes.
with the experimental data, it is not the preferred one for
complexes with symmetric ligands. As it will be discussed
in the next section, the stabilization of the SP S = 0
complexes involves stereochemical effects. However, the
electronic origin of this is the induced symmetry lowering
2 2
1 3
of the charge distribution of the 3A2(b2 e ) and 3E(b2 e )
2
1
1
states in D2d2 symmetry. The corresponding dxy dyz dxz
and dxy dyz dxz or dxy dyz dxz configurations relax
1
1
1
1
2
via the D2d1 f D2 f Ci stabilization procedure to the
2
1
2
2
0
2 2
dxz dxy dyz and dxz dxy dyz configuration (notation as
for z//N-Ni-N direction), respectively, stabilizing the
1E ground spin state (Figures 8 and 9).
and it can only exist in the minimum of the b2 e config-
uration. On the other hand, compressed structures will be
1 3
stabilized in the region of the b2 e configurations, show-
In summary, up to this point we have presented evi-
dence, supported by X-ray crystallographic data and
multiconfiguration ab initio calculations, that the inter-
conversion pathway can be treated as a PJT-induced
phenomenon (if the SOC or ligand field influence to the
3E degenerate state are excluded). The results match the
expected isomerization routes applied for elongated or
compressed Td Ni(II) complexes containing mono- and
bidentate ligands.75
ing twisting distortions to an extent that will be controlled
by the interactions between the peripheral R groups. In
that respect, as Figure 7b shows, twisting distortions
toward R < Rt will stabilize the S = 1 state along the
D2d f C2v pathway, as in the case of NiMe,MeL2,23 while
twisting distortions to the opposite direction, R > Rt, will
stabilize the S=1 structures along the D2d f D2 pathway,
9,24
as in the case of 1Td
.
Due to the shallow D2 mini-
mum, under certain circumstances, both stereoisomers
Spin-Orbit Coupling Contributions. Insight to the SOC
contributions to the ground state as a function of the
twisting angle R2 is provided by the calculations of the
SOC spin states presented in the Supporting Information,
Figure S5. Although mS is not a good quantum number,
we will refer to the states as |0æ and |(1æ, respectively. In
accordance to the magnetic measurements on the D2d
limit of structural conformations (twisting angle range
(Td twist and SP trans-twist) can be observed, as in the
9,24
,
case of 1Td
1 ,
28 and Ni[iPr2P(Se)NP(Se)iPr2]2.32
SP
On the other hand, in the presence of asymmetric
ligands, strong deviations from the Td geometry are
expected due to the intrinsic structural characteristics of
the ligands. This suggestion is supported by the structure
of the asymmetric iPr,PhLK ligand (vide supra), which
upon its coordination to metal ions induces deviations
from the Td structure, as in the case of the MiPr,PhL2 (M =
2 2
108ꢀ<R2<117ꢀ), the ground state, 3A2(b2 e , |0æ) is non-
Mn, Co) complexes.33 For the Ni(II) complexes involving
magnetic. It is important to note that the minor stabi-
3
1 3
2 1 1
asymmetric ligands, the singlet excitation b2 e e
lization energy of the ground spin state E(b2 e , |(1æ)
(dxy f dyz) for larger twisting angles is accompanied with
a strong SOC contribution from the low-lying states
f
2 2 0
b2 e e via spin-relaxation effects (depicted in Figure 9)
becomes allowed, and therefore, the SP S = 0 geometry
might be stabilized, as in the case of the NiMe,PhL29 and 3
complexes.
2 2
2 2
1
2
1
3A2(b2 e , |0æ) and 3A2(b2 e , |(1æ) with dxz dxy dyz
electron configuration. On the contrary, the SOC con-
3
1 3
tribution is weaker to the corresponding E(b2 e , |(1æ)
(dxy f dyz), leading to the strong stabilization of the S = 1
state for smaller twisting angles. It is, thus, evident that
Spin Density Delocalization and Covalency. The calcu-
lated electron and spin populations at the BP86 level for
the triplet and singlet ground states in D2d and SP(Ci)
geometries reflect a strongly covalent NiS4 site, with
significant differences observed between the two cases.
The bonding in the D2d case is ‘traditional’ with the metal
d-orbitals being at higher energy than the ligand orbitals.
In the SP(Ci) case, an ‘inverted bonding’ description is
the SOC contribution via PJTE (3A2 þ E) ꢀ e of the 3E
3
3
(high-spin) excited state mixing with the A2 (high-spin)
ground state largely affects the ground-state configura-
tion, dominating the D2d f D2 or the D2d f C2v pathways
at large distorting angles, as discussed above. However,

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5091
Table 3. The Calculated d-d and LMCT Triplet-Triplet Vertical Excitation Energies of Complex 2a
ground state
state (transition)
DDCI2(12,7)
SORCI(12,7)
expt
3A2f
3E(1b2 f 2e)
3019
3027
4567
4569
9371
3157
3201
5043
5196
8719
-
-
3E(1a1 f 2e)
-
-
8500
3B (1b11b2 f 2e)
9439
9439
9719
8719
8720
8991
3A2(1b1,1a1 f 2e)
11 094
11 415
11 549
11 550
12 127
13 039
15 817
16 263
16 267
17 103
17 122
17 569
17 586
17 590
17 825
21 203
21 203
21 241
10558
10 879
11 003
11 004
12 242
12 377
15 141
16 004
16 125
17 070
17 090
17 364
17 396
17 502
17 659
19 546
19 546
19 578
10 200
spinflip
12 127
13 039
15 817
16 263
16 267
17 042
12 242
12 377
15 141
16 004
16 125
17 022
∼12 800-16 100
3E(1b1 f 2e)
17 066
17 070
3A2(1a11b2 f 2e)
LMCT(1e f 2e)
20 893
19 272
∼20 500b
25 134-26 476
28 850-29 468
∼21 000-29 000
a In cm-1, relative to the lowest 3A2 state. Their assignment is based on MR-DDCI2 and SORCI calculations. For the experimentally important d-d
transitions, the SOC corrected energies are also presented (second subcolumn). The calculations were performed on a model structure of NiiPr,iPrL2 in
which the iPr peripheral groups were replaced by H and were optimized at the BP86 level for their positions. b The transition 3A2(3T1, F) f 3A2(3T1, P)
likely overlaps with LMCT and or ligand-based transitions.
obtained in which the ligand orbitals are situated at
higher energy than the metal d orbitals. This behavior
has been reported previously for Ni bis(dithiolene) SP
complexes.80
highly favorable, providing an efficient pathway for ligand-
to-metal charge donation and thus resulting in a highly
covalent Ni-S bond.
Excited States. In this section, a full interpretation of
the absorption spectrum of complex 2 is performed by
MR-DDCI2 and SORCI dynamic correlated methods, tak-
ing into account the SOC interactions. The calculations con-
firm the experimental observations, predicting that the bands
in the visible region of the spectrum, shown in Figure 2 and
Table 3, are composed from 3A2 f 3B(1b2,1a1 f 2e,3F), 3A2
f 3A2(1b11b2 f 2e,3F), 3A2 f 3E(1b1 f 2e,3P), and 3A2 f
3A2(1a11b2 f 2e,3P) “d-d” excitations, respectively.
As discussed in the Experimental Section, the room-
temperature experimental spectra of 1Td and 2 reveal
multicomponent bands being traditionally assigned as
3A2 f 3T1(3P) excitations, that are split by low-symmetry
and SOC effects.61 Spin-forbidden transitions are calcu-
€
The Lowdin population analysis from the BP86 calcula-
tions provide 1.5 unpaired electrons on the Ni center and
0.5 unpaired electrons on the sulfur ligands, indicating
significant S f Ni charge donation. This can also be seen
in thequasirestricted molecular orbital diagram in Figure3.
The dyz and dxz MOs will be involved in covalent interac-
tions with the corresponding S lone pair (LP) orbitals,
providing σ-covalent Ni-S bonding interactions. In addi-
tion, a more complete picture is revealed from the localized
β spin S LPs which are distributed in the Ni-S-P sequence
as 25% Ni, 70% S, and 5% P. Thus, this data indicates
significant spin delocalization from the metal to the ligand
via a π-pathway. This trend is also reflected in the values of
0
the P-N-P angles of the Td and SP NiR,R L2 complexes
(Supporting Information, Table S3) and supported by
experimental data, such as the 31P NMR paramagnetic
shifts (Table 1) as well as the orbital reduction factor k
(Table 2) determined by the analysis of magnetic properties
(vide supra). In the SP NiH,HL2 model, the decrease of the
S-Ni-S bite angle to 98ꢀ results in better alignment of the
S-3p orbitals with respect to the Ni-3dxy orbital, increas-
lated very close to the A2 f E (1b1f2e,3P) transitions,
thus they might gain intensity and contribute to the shape of
the multicomponent band. In fact, an average of a non-neg-
ligible borrowing intensity with an oscillator strength 6.0 ꢀ
103 is calculated for these spin-forbidden transitions.
The 3A2 f 3A2 (1a11b2 f 2e,3P) excitation is predicted
at ∼19 000 cm-1, with very low intensity. In the experi-
mental spectrum, it is presumably obscured by CT excita-
tions, as can be seen in the Gaussian resolution analysis in
Figure 2. In the CT region, the first three LMCT triplet-
triplet vertical excitations originate to the 1e f 2e electron
promotions (Figure 3). The most intense band with an os-
3
3
€
ing their mixing, as reflected in the Lowdin population
analysis of the Ni dxy-S 3p (50.4 and 49.6%, respectively),
thus showing a stronger Ni-S σ-bond relative to the Td
isomers. In fact, the lowest unoccupied molecular orbital
(LUMO) is the antibonding combination of the Ni-3dxy
and the ligand σ in-plane-out-of-phase orbitals. Due to the
ligand geometry, the overlap between these two orbitals is
cillator strength of 2.1 ꢀ 106 is calculated at 28 850 cm-1
,
in very good agreement with the experimental spectrum.
Magnetic Properties Calculation. The 45 microstates
that span the d8 ligand field manifold will split under
the action of the low-symmetry crystal field and the SOC.
The SOC states can be classified according to double
(80) Szilagyi, R. K.; Lim, B. S.; Glaser, T.; Holm, R. H.; Hedman, B.;
Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2003, 125, 9158–9169.

5092 Inorganic Chemistry, Vol. 49, No. 11, 2010
Maganas et al.
group symmetry, as shown in the Supporting Informa-
tion, Figure S2. In the case of tetragonally elongated
structures, the ZFS has its usual meaning, and one
obtains large and positive D values (εΓ5 - εΓ1), consistent
with the nonmagnetic behavior observed experimentally.
Although the mS is not a good quantum number, we will
refer to the states as |0æ and |(1æ, respectively.
In general, multiconfigurational ab initio calculations
are more suitable in the presence of near orbital degen-
eracy than in DFT methods because they can represent all
magnetic sublevels explicitly and on an equal footing.
Thus, the combination with the quasidegenerate pertur-
bation theory is suitable to attack problems of the kind
met here. In the QDPT approach, the SOC (and SS)
effects are treated through diagonalization of the SOC
(and SS) operators, in the basis of the triplet and singlet
roots obtained from SA-CASSCF or MR-DDCI or
SORCI calculations.
Attempts to experimentally observe EPR signals of the
0
S=1 Td NiR,R L2 complexes proved unsuccessful. Com-
plex 1Td remained EPR silent up to extremely high-
magnetic fields (15 T) and frequencies (600 GHz).81
Therefore, the ZFS in this system is expected to exceed
25 cm-1. In the following, we provide a systematic inter-
pretation of the SH parameters of such systems from the
point of view of quantum chemical calculations. The
findings are compared with the experimental magnetic
measurements discussed above.
Anisotropic covalency is introduced at the level of the
molecular orbitals, from which the multiconfigurational
states are constructed. Unfortunately, Hartree-Fock-
based methods, such as CASSCF, will strongly under-
estimate metal-ligand mixing and, hence, underestimate
anisotropic covalency. This shortcoming is very impor-
tant since the estimated covalency from magnetic
measurements is large (k ∼ 0.35-0.5). However, in corre-
lated ab initio MR-DDCI or SORCI calculations
employed on top of the CASSCF wave functions, these
shortcomings are considerably reduced and, therefore,
realistic predictions of the ZFS are expected. In addition
to the second-order perturbation theory equations for the
ZFS, ab initio methods treat the triplet and singlet states
in infinitive order. We can, therefore, extract the ZFS
parameter D by using the exact solution of the S = 1 SH
problem for the ZFS. In fact, for the axially elongated
structure at twisting angle R2 = 110, the MR-DDCI
calculation up to the second-order perturbation theory
predicts a value of D = 73 cm-1 (E/D = 0.07). Further-
more, calculation of the eigenvalues of the diagonalized
3 ꢀ 3 ZFS matrix gives the energy separation between the
Spin Hamiltonian Parameters: g tensor, ZFS. In order
to interpret the g and D tensors of the D2d elongated NiS4
core, it is important to know the selection rules of the SOC
operator. Under D2d symmetry, the z component trans-
3
forms as B and the x,y components as E. Thus, the B
state has nonzero SOC with B states via the z component
as well as with E states via the x,y components. The g
tensor calculated by the B3LYP method shows, as ex-
pected, small deviations from the free-electron value
along the S4 axis, which coincides with the N-Ni-N
direction (g|| = 2.0031) and significantly large values
perpendicular to it (g^=2.367). These results suggest that
the large positive Δg^ is dominated by contributions from
3
the E(1b2 f 2e) state and much less by contributions
from 3E(1b1 f 2e) because the former represents a
transition from a doubly to a singly occupied orbital,
which is associated with positive g shifts.82,83
two ground spin states as D = (εΓ5- εΓ1) = 60 cm-1
.
Quantitative predictions of the ZFS tensor to a first
approximation were performed with the recently pro-
posed linear response DFT theory.71 The B3LYP/TZVP
coupled-perturbed CP method is used for the calculation
of the ZFS values, which uses revised prefactors for the
spin-flip terms and solves a set of coupled-perturbed
equations for the SOC perturbation, as described else-
where in details.84 The Dss part accounting for the spin-
spin interactions to the ZFS is treated with the ‘UNO’
option which allows the calculation of the SS term with
a restricted spin density obtained from the singly occu-
pied unrestricted natural orbitals. By employing the
above method, we arrive to the following ZFS values:
D=þ58 cm-1 and E/D=0. The origin of the ZFS arises
particularly from spin-orbit coupling (Dsoc = 56 cm-1),
originating from spin-conserving excitations to the SO-
MOs 3E(2b1 f 2e), 3E(1a f 2e), and 3E(1b2f2e), while
only 3% involves spin-flip forbidden excitations. The
spin-spin interactions account only for ∼4% (3 cm-1) of
the D value. Thus, these results are surprisingly good,
given that in the presence of near orbital degeneracy, all
quantum chemical methods have difficulties to predict
accurate values of the SH parameters.
We have furthermore checked the consistency of the
DFT prediction method with well established semiquan-
titative ligand field methods, by employing standard
ligand field arguments combined with excited state en-
ergies calculated by correlated ab initio methods.82 The
computations indicate that there are major contributions
from excitations, which occur within the triplet state. The
final value of D = 55 cm-1 is fairly realistic for tetra-
hedraly coordinated d8 high-spin complexes and is con-
sistent with the DFT and the multireference ab initio
calculated values of D = 57 and 60 cm-1, respectively. In
an effort to check the reliability of the calculated ZFS as
well as the gz and gxy values, we performed a reconstruc-
tion of the experimental magnetic susceptibility compo-
nents by using the closed form formulas for the suscep-
tibility components of an axial ZFS system69,70 (S = 1,
E = 0). Analytical expressions of the equations are
provided in the Supporting Information. The fitting to
the experimental data for NiiPr,iPrL2 is fairly satisfactory
(Figure 4c).
Conclusions
In this work, the synthesis and thorough characterization
of a NiS4-containing complex, NiiPr,PhL2 (3), bearing the
asymmetric dithioimidodiphosphinate iPr,PhL- ligand, is de-
scribed. Complex 3 completes the series of Ni(II) complexes
involving the respective symmetric ligands, namely NiPh,PhL2
(1) and NiiPr,iPrL2 (2). The nature of the R peripheral group
(81) Krzystek, J.; Ferentinos, E.; Maganas, D.; Kyritsis, P. Unpublished
results.
(82) Neese, F.; Solomon, E. I. Magnetoscience - From Molecules to
Materials; Miller, J. S., Drillon, M. Ed.; 2003; Vol. IV, p 345.
(83) Neese, F.; Solomon, E. I. J. Am. Chem. Soc. 1998, 120, 12829–12848.
(84) Neese, F. J. Chem. Phys. 2007, 127, article number: 164112.

Article
Inorganic Chemistry, Vol. 49, No. 11, 2010 5093
i
(Ph or Pr) in defining the NiS4 core geometry and, conse-
quently, the corresponding ground spin state, were shown to
be crucial. Crystallographic data and spectroscopic studies in
the solid state and in solution have proved that the SP NiS4
geometry is observed under specific conditions, only in the
solid state (complexes 1SP and 3), whereas in solution there is
clear preference for the Td geometry of all three, 1-3, com-
plexes. The above observations are confirmed by theoretical
calculations which show that a symmetric R,RL- ligand
environment would tend to stabilize a Td NiS4 structure as
six-membered Ni-S-P-N-C-S chelating rings have so far
been shown to be exclusively SP.87-92
Of special importance is the experimental and theoretical
evidence, presented herein, for a strong spin-orbit coupling
in the paramagnetic Td NiS4 complexes studied, leading to
large ZFS values, which will be a challenge to confirm by
appropriate experimental methods as EPR,34,93,94 MCD,93
and Far-IR magnetic spectroscopy.95
Collectively, our work can form the basis of investigating
structure-properties relationships in other NiS4-containing
systems.
3
2 2
well as the S = 1, A2(b2 e ), ground spin state, against the
expectations stemming from ligand field theory arguments.
This state, however, might be unstable due to the contribution
Acknowledgment. We thank the University of Edin-
burgh, School of Chemistry crystallography service for
data collection and structure solution of complex 3 and
Frank Mabbs, University of Manchester for useful
discussions on the magnetic simulation. D.M. and F.N.
acknowledge financial support from the DFG priority
program SFB 813 “Chemistry at Spin Centers”. P.K.
thanks the Special Account of the University of Athens
(grant 7575) and the Empirikion Foundation for funding
as well as Prof. Georgios Pneumatikakis for his support.
The referees of the manuscript are thanked for their
constructive comments.
of the 3E(b2 e ) JT-active excited states, which induce a strong
1 3
orbital mixing and initiate the Td f SP interconversion path-
way, in terms of a D2d f D2 f Ci symmetric sequence.
Such orbital mixing can be enhanced in complexes with
0
asymmetric R,R L- ligand environments, thus leading to a
SP NiS4-core geometry and a singlet ground state, as has been
experimentally observed for NiMe,PhL2 and complex 3.
In addition, detailed SQUID measurements on oriented
crystals of complexes 1 and 2 are consistent with extensive
delocalization of spin density toward the ligands, a finding with
wider implications concerning the nature of Ni-S bonds in
nickel enzymes,3-5 despite the nonbiological nature of the
dithioimidodiphosphinate ligands. Along these lines, previous
investigations on the covalency of Ni-S bonds have employed
sulfur K-edge X-ray spectroscopic studies on both Td and SP
NiS4- and NiS2N2-containing complexes, respectively, but in
this case, the ligands employed were of different types.85,86
The findings of our work, combined with earlier investiga-
tions,9,28,32 confirm that dichalcogenated (S or Se) imidodipho-
sphinate types of ligands can afford either Td or SP complexes
based either on different R peripheral groups or, simply, on
different experimental conditions. The versatile coordinating
properties of the ligands employed in this work toward Ni(II)
are evident, since the analogous NiS4 complexes containing
Supporting Information Available: Table S1, crystallographic
data for iPr,PhLK and complex 3; Table S2, selected bond lengths
and angles of iPr,PhLH, iPr,PhLK, and 3b; Table S3 a and
0
b selected bond lengths and angles of SP and Td NiR,R L2
complexes; Table S4, IR data; Figure S1, electronic spectra
of 1Td and 3; Figure S2, energy level diagram for four-coordinate
high-spin Ni(II) complexes, and relevant analysis; Figure S3,
31P-{1H} NMR spectra; Figure S4, the most important
NiH,HL2 structures, and relevant analysis; Figure S5, the
2 2
1 3
3A2(b2 e ) and 3E(b2 e ) SOC spin states as well as the LF-ZFS
model equations and magnetic susceptibility equations.
This information is available free of charge via the Internet at
http://pubs.acs.org.
(91) Safin, D. A.; Sokolov, F. D.; Gimadiev, T. R.; Brusko, V. V.;
Babashkina, M. G.; Chubukaeva, D. R.; Krivolapov, D. B.; Litvinov,
I. A. Z. Anorg. Allg. Chem. 2008, 634, 967–971.
(92) Safin, D. A.; Sokolov, F. D.; Szyrwiel, Ł.; Babashkina, M. G.;
Gimadiev, T. R.; Hahn, F. E.; Kozlowski, H.; Krivolapov, D. B.; Litvinov,
I. A. Polyhedron 2008, 27, 2271–2276.
(85) Williams, K. R.; Hedman, B.; Hodgson, K. O.; Solomon, E. I. Inorg.
Chim. Acta 1997, 263, 315–321.
(86) Szilagyi, R. K.; Bryngelson, P. A.; Maroney, M. J.; Hedman, B.;
Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2004, 126, 3018–3019.
(87) Iwamoto, T.; Ebina, F.; Nakazawa, H.; Nakatsuka, C. Bull. Chem.
Soc. Jpn. 1979, 52, 1857–1858.
(88) Zak, Z.; Glowiak, T.; Chau, N. T. T.; Herrmann, E. Z. Anorg. Allg.
Chem. 1990, 586, 136–140.
(89) Sokolov, F. D.; Zabirov, N. G.; Brusko, V. V.; Krivolapov, D. B.;
Litvinov, I. A. Mendeleev Commun. 2003, 72–73.
(90) Brusko, V. V.; Rakhmatullin, A. I.; Shtyrlin, V. G.; Krivolapov,
D. B.; Litvinov, I. A.; Zabirov, N. G. Polyhedron 2006, 25, 1433–1440.
(93) Krzystek, J.; Zvyagin, S. A.; Ozarowski, A.; Fiedler, A. T.; Brunold,
T. C.; Telser, J. J. Am. Chem. Soc. 2004, 126, 2148–2155.
(94) Krzystek, J.; Swenson, D. C.; Zvyagin, S. A.; Smirnov, D.;
Ozarowski, A.; T., C.; Telser, J. J. Am. Chem. Soc. 2010, 132, 5241–5253.
(95) Vongtragool, S.; Gorshunov, B.; Dressel, M.; Krzystek, J.; Eichhorn,
D. M.; Telser, J. Inorg. Chem. 2003, 42, 1788–1790.